Abstract:
|
Objectives: Clinical trial results depend on methods to support results, so reporting should contain methodological criteria. Methods: 24 trials Cochrane Database of Systematic Reviews with "lupus" in the title were selected. One reviewer scored trials using methodological criteria for design, allocation, blinding, reporting and imputation. Results: Most provided baseline data; had 2 arms; provided p values, had missing data, few showed how handled. Fewer claimed double blinded. Also fewer used computer randomization or random number tables. For even fewer: blinding therapy, therapist, patients, outcome assessor, care givers, data analyst, manuscript writer; generation of random numbers, used concealed randomization list, adjusted for baseline differences, stratified patients; sample size justification, confidence intervals, statistical software; used blocking; had CONSORT flowchart, stated randomization integrity, checked assumptions an specified subgroups, gave a randomization audit trail, Numbers Needed to Treat and provided analysis computer. Conclusions: Cochrane lupus trials report few methodological criteria. CONSORT checks should help make lupus trials more credible.
|
Copyright © American Statistical Association.